This is a ros pkg for gazebo actor plugin.
- Ubuntu 16.04
- ROS-kinetic
- Gazebo 8 (with actor suport)
- python-lxml
- turtlebot3
- turtlebot3_msgs
- turtlebot3_simulations
-
Add the repositories of Gazebo 8 and ROS kinetic
-
Install Gazebo 8, Ros kinetic in buntu 16.04 and other dependencies.
sudo apt-get install ros-kinetic-desktop-full
sudo apt-get install ros-kinetic-gazebo8-ros-pkgs
- Build packages
cd /path/to/workspace/src
git clone [email protected]:onlytailei/gym_ped_sim.git
catkin build
roslaunch turtlebot3_social default.launch
-
actor_plugin
Build based on a Gazebo official example. This node broadcasts the tf of every actor. Social force model is applied in every actor to interactive with each other. -
actor_services
The python files help to create several gazebo sdf files quickly. There is a rviz file for visualization. -
turtlebo3_social
In our socially compliant pedestrian simulator, we collect data by mounting a depth sensor onto one of the pedestrians, to the height matching that of real-world setups. Then, the social force model, as described in the paper, is used to label each incoming depth image with their corresponding social force. -
data_collection
To save the related dataset.
The collected pedestrian navigation dataset contains:
- depth image
- RGB image
- target
- social force classification
- social force
- sum force
Please reference gym_style_gazebo
This is the reference implementation of the plugins and for the paper Socially-compliant Navigation through Raw Depth Inputs with Generative Adversarial Imitation Learning. If it helps your research, please cite:
@inproceedings{tai2018social,
author={L. Tai and J. Zhang and M. Liu and W. Burgard},
booktitle={2018 IEEE International Conference on Robotics and Automation (ICRA)},
title={Socially Compliant Navigation Through Raw Depth Inputs with Generative Adversarial Imitation Learning},
year={2018},
pages={1111-1117},
doi={10.1109/ICRA.2018.8460968},
ISSN={2577-087X},
month={May},
}