The ROS Gazebo Gym framework integrates ROS and Gazebo with gymnasium to facilitate the development and training of RL algorithms in realistic robot simulations. It comes equipped with several ready-to-use simulation environments, allowing for a diverse range of applications and experimentation. This framework provides a streamlined way to apply simulation-trained algorithms to actual robots, thereby enhancing their real-world applicability. Our goal with ROS Gazebo Gym is to establish a robust foundation for RL research in robotics, aiding in the advancement of practical, real-world robot control."
Please see the docs for installation and usage instructions.
We use husky pre-commit hooks and github actions to enforce high code quality. Please check the CONTRIBUTING.md before contributing to this repository.
Note
We used husky instead of pre-commit, which is more commonly used with Python projects. This was done because only some tools we wanted to use were possible to integrate the Please feel free to open a PR if you want to switch to pre-commit if this is no longer the case.