/Action-Based-Representation-Learning

Primary LanguagePython

Action-Based Representation Learning for Autonomous Driving

This repository is forked from code associated with the paper: Action-Based Representation Learning for Autonomous Driving

Experiments Summary

The processes can be defined as four types:

  • Train an encoder model (Behaviour Cloning (BC), Inverse, Forward, ST-DIM)
  • Train a MLP for affordances outputs. The pre-trained encoder model will be used.
  • Validation on affordances prediction.
  • Actual drive using controller tuned with affordances prediction.

Setting Environments & Getting Datasets

  1. Download the dataset customized.

  2. change path to your dataset folder with SRL_DATASET_PATH in run_customized.sh and run_analysis.sh:

  3. Download the repository

        git clone https://github.com/AIasd/Action-Based-Representation-Learning.git

  1. change path to your downloaded repository with ACTIONDIR in run_customized.sh and run_analysis.sh:

  2. Download the CARLA version we used with this link, and put it inside your downloaded repository folder

  3. change the absolute path file_dir in carl/cexp/env/environment.py to your local folder path.

Training Encoder

  1. Define configuration files for training. Refer to files in configs folder

  2. Run the run_customized.sh file with "train_encoder" process by only commenting in the corresponding line: python3 main.py --single-process train_encoder --gpus 0 --encoder-folder ENCODER --encoder-exp customized where --single-process defines the process type, --gpus defines the gpu to be used, --encoder-folder is the experiment folder you defined in config folder, and --encoder-exp is the experiment you defined inside the experiment folder.

Training MLP for affordances

  1. Define configuration files for training. Refer to files in configs folder

  2. Run the run_customized.sh file with "train_encoder" process by only commenting in the corresponding line: python3 main.py --single-process train --gpus 0 --encoder-folder ENCODER --encoder-exp customized --encoder-checkpoint 30000 -f EXP -e customized

Validate on affordances prediction

  1. Run the run_customized.sh file with "train_encoder" process by only commenting in the corresponding line: python3 main.py --single-process validation --gpus 0 --encoder-folder ENCODER --encoder-exp customized --encoder-checkpoint 30000 -f EXP -e customized -vj $ACTIONDIR/carl/database/CoRL2020/customized.json

Video

video

Reference:

  • Yi Xiao, Felipe Codevilla, Christopher Pal, Antonio M. Lopez, Action-Based Representation Learning for Autonomous Driving.

     @article{Xiao2020ActionBasedRL,
 title={Action-Based Representation Learning for Autonomous Driving},
 author={Y. Xiao and Felipe Codevilla and C. Pal and Antonio M. L{\'o}pez},
 journal={ArXiv},
 year={2020},
 volume={abs/2008.09417}
 }

  • Coiltraine, which can be used to easily train and manage the trainings of imitation learning networks jointly with evaluations on the CARLA simulator.

  • Cexp, which is a interface to the CARLA simulator and the scenario runner to produce fully usable environments.