DeepSim Toolkit is an open-source reinforcement learning environment build toolkit for ROS and Gazebo. It provides the building blocks of extensive advanced features such as collision detection, behavior controls, domain randomization, spawner, and many more to build complex and challenging custom reinforcement learning environment in ROS and Gazebo simulation environment with Python language.
DeepSim whitepaper is available at https://arxiv.org/abs/2205.08034. If you reference or use DeepSim in your research, please cite:
@misc{la2022deepsim,
title={DeepSim: A Reinforcement Learning Environment Build Toolkit for ROS and Gazebo},
author={Woong Gyu La and Lingjie Kong and Sunil Muralidhara and Pratik Nichat},
year={2022},
eprint={2205.08034},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
The source code is released under Apache 2.0.