The tensorflow implmentation of paper: Learn to Steer through Deep Reinforcement Learning
Videos of our experiments are available at Video1 and Video2.
Python 3.5
Tensorflow 1.13.1
ROS Kinetic
Clone the pioneer_utils folder to your catkin workspace and catkin_make:
cd ~/catkin_ws/
catkin_make
source ~/catkin_ws/devel/setup.bash
Launch the simulator with command:
roslaunch pioneer_utils p3_world.launch
Start training with command:
python main.py
If you find this method useful in your research, please cite:
@article{wu2018learn,
title={Learn to steer through deep reinforcement learning},
author={Wu, Keyu and Abolfazli Esfahani, Mahdi and Yuan, Shenghai and Wang, Han},
journal={Sensors},
volume={18},
number={11},
pages={3650},
year={2018},
publisher={Multidisciplinary Digital Publishing Institute}
}