highway_irl

"highway_task" folder contains the task. Follow the instructions in the readme file in the folder to install and run the task.

Analysis code description:

1_install_env: install required packages (requires setuptools==62.6.0 & wheel==0.38.0).

2_dqn_train_exp_iter: trains a DQN model

3_dqn_collect_batch: generate behaviors using a DQN model

4_run_highway_AIRL: AIRL for the data from DQN

5_data_to_traj: generate state-action trajectories from the task data

6_run_AIRL_data: AIRL for the data from experiments

descriptive_analysis_final: shows the correlation between BIS and behavioral task measures

IRL_group: shows the model fit of the AIRL models

IRL_raw_reward: illustrates the IRL rewards in simplified state spaces

reward_trajectory: illustrates the IRL reward trajectories for overtaking and crashing

Name		Name	Last commit message	Last commit date
Latest commit History 4 Commits
.ipynb_checkpoints		.ipynb_checkpoints
AIRL		AIRL
behavior_v2		behavior_v2
data_npy		data_npy
dqn		dqn
dqn_traj/35quick120_512		dqn_traj/35quick120_512
env_configure		env_configure
highway_task		highway_task
shared		shared
trained_models_AIRL		trained_models_AIRL
traj_from_data_v2		traj_from_data_v2
1_install_env.ipynb		1_install_env.ipynb
2_dqn_train_exp_iter.py		2_dqn_train_exp_iter.py
3_dqn_collect_batch.py		3_dqn_collect_batch.py
4_run_highway_AIRL.py		4_run_highway_AIRL.py
5_data_to_traj.py		5_data_to_traj.py
6_run_AIRL_data.py		6_run_AIRL_data.py
IRL_group.ipynb		IRL_group.ipynb
IRL_raw_reward.ipynb		IRL_raw_reward.ipynb
README.md		README.md
descriptive_analysis_final.ipynb		descriptive_analysis_final.ipynb
reward_trajectory.ipynb		reward_trajectory.ipynb