Torcs Behavior Cloning

Learn to drive in TORCS from expert example.

How to use

Python dependencies

• Install anaconda python3 (https://www.anaconda.com/download/#linux)
• clone this repo.
• cd into the gym-torcs directory.
• create a virtual environment with Conda, Anaconda's CLI tool: conda env create -f environment.yml -n sdcc. If you get an error at the end, don't worry about it. (TODO: FIX)
• activate the environment source activate sdcc
• install remaining packages through pip pip install -r requirements.txt

Dependencies for the Simulator

• x-automation7: sudo apt install xautomation
• other dependencies: sudo apt install libglib2.0-dev libgl1-mesa-dev libglu1-mesa-dev freeglut3-dev libplib-dev libopenal-dev libalut-dev libxi-dev libxmu-dev libxrender-dev libxrandr-dev libpng12-dev ffmpeg

Configuring and installing the simulator

• cd into gym-torcs/vtorcs-RL-color
• ./configure
• make
• sudo make install
• sudo make datainstall
• Go back up a directory (to gym-torcs).
• Configure permissions with sudo ./configure_permissions

Recording Images with the simulator

• cd to gym-torcs directory -python snakeoil3_gym.py
• That will record images in to ./databases/<track>/<trial>/imgs directory.
• To make images into a movie, use fmpeg on the directory: Ffmpeg -i <path>/imgs/%05d.png video.webm

Training the agent

• be in gym-torcs directory.
• assumes you have already recorded some data somewhere.
• python main.py -d ./databases/<track>/<trial>