What is gym-ignition?

gym-ignition is a framework to create reproducible robotics environments for reinforcement learning research.

The aims of the project are the following:

  • Provide unified APIs for interfacing with both simulated and real robots.

  • Implement the simulation backend interfacing with the Ignition Gazebo simulator.

  • Enable a seamless switch of all the physics engines supported by Ignition Gazebo.

  • Guarantee the reproducibility and the scalability of the simulations by using the simulator as a library, without relying on any network transport.

  • Simplify the development of OpenAI Gym environments for robot learning research.

gym-ignition targets both control and robot learning research domains:

  • Researchers in robotics and control can simulate their robots with familiar tools like Gazebo and URDF, without the need to rely on any middleware.

  • Researchers in robot learning can quickly develop new robotic environments that can scale to hundreds of parallel instances.

To know more about why we started developing gym-ignition, why we selected Ignition Gazebo for our simulations, and what are our long-term goals, visit the Motivations page.

We are building an entire ecosystem around gym-ignition, if you’re interested have a look to the other projects:

ScenarI/O and gym_ignition

Robot Models

Ignition Plugins