UAV Obstacle Avoidance using Deep Recurrent Reinforcement Learning with Temporal Attention

The code is implemented in Tensorflow(version = 1.1.0) and Keras.

Requirements

The code is based on Python 2. Install dependency by running:

pip install --user -r requirements.txt

How to run

There are two types of DQN implementation with gpu: Keras and Tensorflow.
You can choose different implementation by altering line 15 in main.py

Train original DQN:

python main.py --task_name 'DQN'

Train Double DQN:

python main.py --ddqn --task_name 'Double_DQN'

Train Dueling DQN:

python main.py --net_mode=duel --task_name 'Dueling_DQN'

Train Recurrent DQN:

python main.py --num_frames 10 --recurrent --task_name 'Recurrent_DQN'

Train Recurrent Temporal Attention DQN: (Using dqn_tf_temporalAt.py by uncommenting line 18 in main.py)

python main.py --num_frames 10 --recurrent --a_t --selector --task_name 'TemporalAt_DQN'

Train Recurrent Spatial Attention DQN: (Using dqn_tf_spatialAt.py by uncommenting line 21 in main.py)

python main.py --num_frames 10 --recurrent --a_t --selector --task_name 'SpatialAt_DQN'

Test trained model (e.g. Spatial Attention DQN):

python main.py --num_frames 10 --recurrent --a_t --selector --test \
--load_network --load_network_path=PATH_TO_NET

Acknowledgement

This code repository is highly inspired from work of Rui Zhu et al link.