songanz/highway-env-agents

highway-env&agents

A collection of environments for autonomous driving and RL agents for decision making

1. Installation

1. Builder docker using docker file

cd docker
./build_docker.sh

2. After successfully build the docker

in docker folder, run:

./run_container.sh

2. The environments

Please check the readme of sub module highway-env

3. The RL agents

This repo include two popular RL library as sub modules:

1. stable_baselines3
2. rl_agents

Please refer to the readme of the original sub modules

4. Examples

4.1 Simple Usage of highway-env

import sys
import os
highway_env_path = os.path.join(os.getcwd(), "highway-env")
sys.path.append(highway_env_path)

import gym
import highway_env

env = gym.make("highway-v0")
obs = env.reset()

done = False
while not done:
    action = env.action_space.sample() # for simple usage, you can just sample action space
    obs, reward, done, info = env.step(action)
    env.render()

4.2 Training

4.2.1 stable_baselines3

Please check script train_stable_baselines3.py which is setup using stable_baselines3

4.2.2 rl_agents

Please check bash script train_rl_agents.sh for setting up training using rl_agents

4.3 Plot training curve

4.3.1 stable_baselines3

Please check script plot.py, and write your own

4.3.2 rl_agents

TBD

4.4 Load trained policy

4.4.1 stable_baselines3

Please check script ani.py, and write your own

4.4.2 rl_agents

TBD

Reference

Gym Documentation