search

AcutronicRobotics / gym-gazebo2

gym-gazebo2 is a toolkit for developing and comparing reinforcement learning algorithms using ROS 2 and Gazebo
https://acutronicrobotics.com
Apache License 2.0
416 stars 107 forks source link

Multi-agent simulation #15

Closed ismorphism closed 5 years ago

ismorphism commented 5 years ago

Hi, everyone, Do you know ways how to simulate multiple robots here in your library or maybe you know other libraries to do it? Could you recommend something?

nzlz commented 5 years ago

Could you be more precise? What is what you want to achieve?

If you want to launch multiple simultaneous instances, yes its possible. We can provide more information if this is what you are looking for.

ismorphism commented 5 years ago

I have the following task - there are 2 TurtleBots, map with obstacles and two target positions (1 robot - 1 target position). And I want to simulate this world with ability to use gym-like interface for design my MARL algorithms. This is exactly my goal.

nzlz commented 5 years ago

This is possible and has been done previously. You should be able to spawn and control both robots in gazebo using different namespaces.

Note that gym-gazebo2 is ROS2 focused and the gym-gazebo ROS (1) version was deprecated, so we don't offer support anymore. Im closing the issue for this reason, however you can post below if you have implementation specific questions.